Plate load tests

1111 Installation ofplates and instrumentation

The plate load tests were performed as a series of tests with plates measuring 0.5 x 0.5, 1 x 1 and 2 x 2 metres, see Chapter 3. The plates were placed in a row with a common centre line. They were placed with the largest plate in the middle and the smaller plates at the ends. The location of the plates and the distances between them were selected in such a way that the influence of one load test on the results of the next test would be negligible.

The reinforced concrete plates were cast in place at the bottom of excavated pits.

The excavation work was started using an excavator down to about 0.3 m above the foundation level. When excavating in loose saturated silt, it can often be observed that an angular distortion occurs in the underlying silt over a depth of more than 0.1 metre. Care was therefore taken to perform this excavation smoothly without excessive deepening of the pit in the separate digging operations. The rest of the excavation was carried out manually. The pits were made square with sides 1 metre longer than the sides of the plates, thus leaving a working space outside the moulds after these had been put in place. The manual excavation was made gradually and plywood sheets were used as platforms for the diggers to stand on in order to protect the carefully excavated base from being disturbed by kneading from the boots. The prefabricated moulds were then lowered in place and all further work was performed from the outside of the moulds, Fig. 4.2.24.

The instrumentation under each of the plates at Vagverket consisted of 3 settlement gauges placed at different depths and one piezometer. The settlement gauges consist of short augers which are screwed down into the soil. The rods are encapsulated in a plastic tube and the annulus is filled with grease. When the auger has been screwed into place, the plastic tube is retracted over a distance corre­

sponding to the intended range of measured settlements. The rods and protecting tubes are extended to pass through the concrete plate.

The settlement gauges were placed close to three corners of the plates in the so called "characteristic point" for stresses and settlements, i.e. at a perpendicular distance of 0.13b from adjacent sides of the plate (where bis the width of the plate).

One gauge for each plate was placed with its centre at a depth of 0.25 metre below the base. The reason for this was to obtain a check of possible excessive deformations because of disturbance of the bottom of the pit in spite of the careful excavation. The other two gauges were placed at depths roughly corresponding to lb and 2b below the plates.

Investigations and load tets in silty soils 85

Starting the excavation by using an excavator.

Finishing the excavation manually.

Fig. 4.2.24 Excavation for the plates in the test field at Vagverket.

SGI Report No 54

86

Settlement gauge.

Placing of moulds, reinforce­

ment and instrumentation.

Concrete plate after casting, with devices for settlement measurement in place. Free ground water can be observed in deeper pockets.

Fig. 4.2.24 Continued.

Investigations and load tets in silty soils 87

The piezometers were installed in order to check that the tests were drained and that all excess pore pressures had dissipated at the end of each load step. The piezometers were electrical and were installed at depths where the pore pressures with consideration to stress increases and drainage paths could be expected to be highest. For the smallest plates, the piezometers were installed in the upper silt layer about 0.8b below the plate and for the largest plate, where significant stress increases were expected to reach down to the clayey layer, the piezometer was placed in the middle of this layer. The piezometers were pushed in using a drill rig on the ground surface at the corners where no settlement gauges were installed and with such inclinations that they would be placed centrally under the plates. After pushing the piezometer into place, the drill rods were retracted, leaving only the piezometer connected to a wire and the signal cable in the ground, Fig. 4.2.25.

Ground anchors with expander bodies

5 10 (m)

I

I

Section A-A

Settlement gauge 0.25 m

Settlement gauge 0.5, 1.75 and 1.0 m respectively Settlement gauge 1.0, 3.3 and 2.0 m respectively ) Pore pressure gauge

Fig. 4.2.25 Layout of the test set up and the instumentation in the test field at Vagverket.

SGI Report No 54

88

Apart from the settlement gauges below the plates, the total settlements of the plates were measured on rods fixed in the concrete at each comer of the plate.

The concrete plates were 0.5 metre thick and reinforced. One day after casting the plates, each mould was raised 0.5 metre and thereby formed a wall over the plate protecting the upper plate surface and the settlement gauges. The space outside the plate and the mould was filled in with excavated soil. This soil was not compacted and thus had a lower density than the original soil. Above the moulds, the in-filled soil was shaped to slope towards the edges of the pits and was thereby prevented from falling down into the moulds , Fig. 4.2.26.

Fig. 4.2.26 Concrete test plate with raised mould and backfilled soil.

Investigations and load tets in silty soils 89

• Reaction system

The reaction system consisted of four ground anchors, two at each end of the row of plates. The ground anchors consisted of expander bodies which had been placed in coarse, medium dense silt at about 17 metres depth. The anchors were connected to the ground surface by tie rods. The capacity of the ground anchors was designed to be such that each of them would safely take the whole reaction load if required due to eccentric loads in the system. However, it was later found that because of repeated loading and unloading the accumulated deformations in the anchor and support system caused a certain tilting and imbalance in the system, which created some problems in the last load test.

The ground surface at each pair of ground anchors was scraped off and levelled.

A number of wooden rafts, "excavator mats", were then stacked perpendicular to the centre line of the plates and adjusted in such a way that the top surfaces of both stacks were horizontal and level with each other. The two railway bridge beams were then placed side by side with the ends resting on the stacks and were then bolted together. Their common centre line was carefully aligned with the centre line of the plate row. Two short beams were then placed on top and across the ends of the large beams and were connected to the tie rods from the ground anchors.

The tie rods were pre-stressed and the whole reaction system was thereby fixed, Fig. 4.2.27.

Fig. 4.2.27 Reaction system being mounted.

SGI Report No 54

90

II Loading system

The loading system is a universal all purpose system developed at SGI and capable of providing monotonic loading, constant load steps, loading in a pre-set pattern, dynamic loading etc. (Bergdahl et al. 1995). The part of the loading system used for the tests at Vagverket consisted of a hydraulic pump and regulation valves operated electronically, a hydraulic jack, a load cell for measuring the load and providing the signal for regulating the hydraulic pressure and a computer for data collection and regulation of the loading process.

For the two smaller plates, a circular stress distribution plate was centered directly on the concrete. The load cell and the jack were then placed on this plate and on top of them further plates were placed as required to fill the gap up to the reaction beams. For the largest plate, an extra square stress distribution steel plate covering a large part of the top surface, was laid in position before the above mentioned assembly was put in place.

When the settlements became excessive and the stroke of the jack insufficient, the load could be released and new gap filling plates added before the loading process was restarted.

II Measuring system

The measuring system consisted of the computerised data collection and load regulation system, the load cell, the piezometers and seven displacement transduc­

ers. The displacement transducers were fixed on a special measurement beam (ladder) placed perpendicular to the reaction system and with the ends resting on levelled ground outside the excavated area. The displacement transducers meas­

ured the movements of the top of the plate and the settlement gauges in relation to this ladder. The whole system was protected from sun and weather by tarpaulins using the reaction beams as a ridge. The measurement beam was supplied with a fixed vertical scale and was continuously levelled by a high precision instrument in order to check that the reference level was stable, Fig. 4.2.28.

The pore pressures in the piezometers just outside the test area and the ground water level in the observation holes inside the area were also monitored continu­

ously during the period of load testing.

Investigations and load tets in silty soils 91

Load test in progress. The measuring system is protected by a tarpaulin and the data collection and load regulation systems are inside the cabin.

A view under the tarpaulin, showing the measurement ladder with the fixed scale and gap filling plates between the reaction system and the hydraulic jack.

Data processing and continuous evaluation of the test during its progress.

Fig. 4.2.28 Load tests in progress.

Data collection and load regulation system.

An intercommunication line is also connected to the test pit.

SGI Report No 54

92

• Loading procedure

The load tests were to be performed in steps with maintained constant loads. In order to ascertain that all excess pore pressures would dissipate during the time for each load step, some preliminary calculations were made using somewhat con­

servative assumptions. On the basis of these calculations, it was tentatively decided that each load step for the smallest plate should be applied for 2 hours. The corresponding times for the middle and the largest plates were 3 hours and 5 hours respectively.

Preliminary calculations of the bearing capacity with different methods had given widely different results. For the first test plate, which was the smallest 0.5 x 0.5 metre plate, it was therefore decided to start very carefully with small load steps of only 10 kN and then, if this was found to be feasible after a number of such steps, increase them to 20 kN.

The maximum design load for the reaction system was 800 kN and for the 1 x 1 metre plate it was decided that a loading sequence with steps of 40 kN up to 200 kN and of 50 kN thereafter would meet the requirement of a sufficient number of load steps up to a possible ultimate failure load. For the largest plate, where a bearing capacity failure was considered as very unlikely, the load steps were chosen as one tenth of the maximum load, i.e. 80 kN.